In today modern society, telecommunication networks play a major role in providing global data and voice communication. Monitoring a network has become important to ensure reliable operation, fault detection, timely mitigation of potentially malicious activities and more. To ensure the vitality of the company, many companies have employed network taps in order to monitor the data traffic flowing through their networks.
To facilitate discussion, FIG. 1 shows a simple diagram of a network environment with a 10/100 megabytes tap. In a typical network environment, data traffic may be flowing between two network devices (Network A and Network B). In an example, data traffic may flow from a port 102 of Network A to a port 104 of Network B. Both port 102 and port 104 may be RJ45 jacks that support Ethernet-over-twisted pairs.
In a 10/100 megabytes network environment, the direction of the data traffic is usually known and uni-directional. In an example, twisted pair pins 1-2 of port 102 and twisted pair pins 1′-2′ of port 104 may be designated as the transmitting pairs while twisted pair pins 3-6 of port 102 and twisted pair pins 3′-6′ of port 104 may be designated as the receiving pairs. For example, data traffic may flow from twisted pair pins 1-2 of port 102 along paths 108/110 to twisted pair pins 3′-6′ of port 104. Similarly, data traffic coming in from port 104 may flow from twisted pair pins 1′-2′ along paths 112/114 to twisted pair pins 3-6.
Since the directional flow of the data traffic within the network environment is known, a network tap 116 may be configured to tap into the paths (108/110 and 112/114) to monitor the data traffic flowing between the two network devices. In an example, data traffic flowing from port 102 to port 104 may be copied and sent along paths 120/122 to a set of physical layer interfaces (PHYs) 118 of network tap 116 before being forwarded to a monitoring device 128. Given that the flow of data traffic is predictable, network tap 116 may be a passive tap. In other words, network tap 116 is not required to be an inline tap. Accordingly, power loss to network tap 116 has substantially little impact on the data traffic (zero delay). In an example, latency and/or data loss may be substantially minimal. Discussion about zero delay on 10/100 megabytes tap is provided in a related application entitled “Zero-Interrupt Network Tap,” filed Apr. 28, 2004 by Matityahu et al. (application Ser. No. 10/834,448), all of which are incorporated herein by reference.
However, in a faster Ethernet environment, such as a gigabit Ethernet, the direction of the data traffic is usually bidirectional and unpredictable. To facilitate discussion, FIG. 2 shows a simple diagram of a network environment with a gigabit tap. Consider the situation wherein, for example, data traffic is flowing between a port 202 of Network A to a port 204 of Network B. Both port 202 and port 204 may be RJ45 jacks that support Ethernet over twisted pairs. To establish a communication link between the two ports, auto-negotiation may be performed. In auto-negotiation, the communication link may be established based on the fastest transmission mode available for the two network devices (such as Network A and Network B) based on common transmission parameters, such as speed of the link and configuration mode (e.g., half-duplex, full-duplex, and the like). Once a communication link is determined, data traffic may then be transmitted between the two network devices.
Given that the direction of the data traffic within a fast Ethernet environment may be unpredictable, an inline tap arrangement may be employed. With an inline tap arrangement, data traffic flowing between port 202 of Network A and port 204 of Network B is configured to flow through a network tap 206. Thus, instead of a communication link established between the two network devices, a communication link may be established between network tap 206 and each of the network devices. In other words, a communication link may be established between Network A and network tap 206 and between Network B and network tap 206. In an example, a network twisted pair pins 1-2 of port 202 may be configured to send data traffic to a tap twisted pair pins 3′-6′ of PHY 208. Upon receiving the data traffic, PHY 208 may then forward the data traffic onward to Network B via a tap twisted pair pins 1-2 to a network twisted pair pins 3′-6′ of port 204 while a copy of the data traffic may be forwarded to a monitoring device 228.
Since the network tap is an inline device, each time the network tap experiences a power disruption (either power is turn on or on the path between Network A and Network B may be renegotiated. In an example, network tap 206 is taken offline for maintenance. When the network tap is taken offline, a new communication link is negotiated to establish a path between Network A and Network B. In an example, a set of relays 210 may be triggered to establish a direct route from Network A to Network B (instead of going through network tap 206).
Each time network tap experiences power disruption, the communication link is lost and a new communication link may have to be established. Consider the situation wherein, for example, network tap 206 is turn off. Thus, the communication links between network tap 206 and Network A and Network B are broken. To enable data traffic to flow between the two network devices, a new communication link may be established. In other words, when the communication links are broken, a set of mechanical relays 210 may be triggered to create a new path. The time required to trigger set of mechanical relays 210 and to enable Network A and Network B to perform auto-negotiation may require a few milliseconds. The latency experienced during this time period may have dire financial consequences. In an example, in a financial industry, a latency of a few milliseconds can result in millions of dollars loss.
Accordingly, arrangements and methods for providing zero delay in a faster Ethernet environment (such as a gigabit Ethernet environment) are desirable.